Light touch contact as a balance aid.

Key Words: Cane, Fingertip, Mobility aid, Neural plasticity, Posture, Rehabilitation, Somatosensory, Vestibular.

Mobility aids such as canes and crutches are commonly used for rehabilitation from musculoskeletal or neuromuscular injuries and for balance problems with elderly individuals. A primary function of such aids is to reduce the risk of fails, because falls may lead not only to physical injury but to loss of confidence and to restrictions on mobility, which may have psychological repercussions (eg, depression). In addition, mobility aids are used to decrease loads on joints or limbs recovering from injury. The literature on mobility aids deals primarily with three general areas(1): 1) the scientific study of variables involved during locomotion with a mobility aid, 2) their therapeutic use as rehabilitation aids, and 3) design modifications. Few studies, however, have systematically investigated these categories for the proper prescription and use of ambulatory aids.(2-4) Most studies have focused primarily on the biomechanical support provided by walking canes or crutches.(5-7) Forces applied to a cane are often measured in relation to forces applied at each foot for comparison of forces applied without cane use. Functional support provided by a cane is often equated with the resulting decrease in force required of the affected limb; the greater the support provided by the cane, the more effective its use. This belief has led to recommendations such as holding a cane in the hand contralateral to the affected limb, because the long lever arm reduces the forces across the hip joint to less than half of that of unaided locomotion.(3)

Assessing the role of the cane as a mobility aid purely in terms of absolute biomechanical forces may be underestimating its potential use in the rehabilitation of patients with balance disorders. There are many properties of cane use that have not been thoroughly investigated. Murray et at,(8) for example, demonstrated that the timing relationships between applied cane forces and the duration of stance were related to the functional use of the cane. Patients with ankle arthropathy applied peak cane force late in the stance phase of the disabled limb, suggesting that the cane was used to push forward. In contrast, patients with degenerative joint disease of the hip applied an initial peak thrust early in the stance phase, suggesting that the cane was used for restraint. Thus, an important consideration in the functional use of a cane may be the timing relationships observed between applied forces through the cane and stance forces for a particular injury or disorder that leads to poor balance.

There are also instances in which it is desirable to limit the physical support derived from a cane. Persons with lower-extremity amputations must learn to gradually shift their weight from a cane to a prosthesis to avoid residual limb irritation at the limb/socket interface.(9) Such patients often have difficulty estimating how much weight to apply to a cane to help support the affected limb. Moreover, Murray et al(8) showed that persons with above-knee amputations applied small cane forces prior to the stance phase with the prosthesis, suggesting that the cane was providing sensory information before the onset of prosthetic weight bearing.

The argument for the importance of sensory input from a balance aid is bolstered by the fact that canes are often prescribed for patients with balance disorders that stem not from an orthopedic problem but from neurological damage. For example, many patients without a functioning vestibular system have poor balance control due to the lack of sensory information about head movement, which is crucial for stable locomotion. (10-12) Patients with Parkinson's disease, whose walking patterns resemble a rigid shuffle, reported increased comfort when using a cane, even though the force applied through the cane was less than that needed for physical support of the body.(8)

The mechanical support provided by a balance aid may be the primary benefit for certain conditions (eg, hip replacement). The spontaneous adoption of different timing strategies with cane use at very low applied force levels, however, emphasizes that patients derive substantial orientation information from a hand-held cane. An underestimated source of support may be the orientation information provided from somatosensory stimulation of the hand and arm through contact of the cane with the ground or a rigid object. Recent investigations(13,14) have shown that contact cues from the fingertip provide information that leads to reduced postural sway in subjects without balance impairments and in patients with bilateral vestibular loss, even when the applied forces are physically inadequate to stabilize the body. Sighted and congenitally blind individuals may use a cane to stabilize upright stance, even at very low force levels, in the same fashion as the fingertip.(15) To provide some background on how touch and pressure cues stabilize upright stance, the results of studies on fingertip contact cane use and the control of quiet upright stance are summarized below.

Light Touch Contact Studies

Light touch contact of a fingertip to a stable surface reduced postural sway in subjects standing on one lower extremity(16) and in a heel-to-toe stance.(13,14) Figure 1 depicts a subject in the heel-to-toe stance (left lower extremity in front of right lower extremity) on a force platform touching a device used to measure the forces applied by the tip of the right index finger. The touch apparatus consisted of a horizontal metal bar attached to a metal stand, parallel with the sagittal plane of the subject. Subjects placed their right index finger on the middle of the bar while strain gauges mounted on the bar transduced the lateral (F.sub.L) and vertical (F.sub.V) forces applied by the fingertip. Subjects were tested with eyes open and closed in three fingertip contact conditions: 1) no contact, during which the subjects' arms hung passively by their sides, 2) touch contact, in which the subjects could apply up to 1 N ([approximately equal to]100 g) of force on the touch apparatus before an auditory tone signaled the threshold of applied force, and 3) force contact, during which subjects could apply as much force as desired. In the light touch condition, if 1 N of force was exceeded, an auditory alarm went off, indicating that the subjects should apply less force without losing contact with the surface. The light touch task is very easy to perform. After just one practice trial to get a "feel" for the threshold force, subjects rarely set off the alarm (<5% of the light touch trials in all the experiments to date). Figure 2 shows the combined results from five subjects.(13) Mediolateral center-of-pressure (CO[P.sub.x]) mean displacement(*) was highest in the eyes closed-no contact condition and reduced in all other conditions. Post hoc tests revealed that touch and force contact lowered mean CO[P.sub.x] displacement equivalently, with or without vision present (no touch, eyes open or closed > light touch and force touch, eyes open or closed) despite mean fingertip force levels that were over 10 times greater with force contact (4 N) than with touch contact (0.4 N).

In a model designed to study the reduction in body sway due to static and dynamic mechanical forces at the fingertip,(16) contact forces of 0.4 N predicted a 2% to 3% reduction of sway. Touch contact, however, reduced sway by 50% to 60% in all subjects. This finding suggests that fingertip forces in the touch contact conditions are inadequate to stabilize upright stance. Subjects must rely on musculature remote from the fingertip to arrest sway toward the touch bar, because touch contact forces alone are not large enough.

Temporal Relationships

When fingertip forces are providing physical support, force levels are expected to increase and decrease with body sway toward and away from the contact surface, respectively. When fingertip contact forces are limited to 1 N, the contact forces can no longer be allowed to rise and fall with body sway without exceeding the threshold force. Thus, a different temporal relationship is expected between contact forces and body sway with light touch. Figure 3 shows the time series of CO[P.sub.x] displacement and [F.sub.L] in typical force contact (Fig. 3a) and touch contact (Fig. 3b) trials. Maximum cross-correlations and their respective time lags are also shown. Correlations between CO[P.sub.x] displacement and [F.sub.L] were highest with force contact ([rho][approximately equal to]0.9), with very small time lags between the two signals (<50 milliseconds).

This finding means that contact forces in the force contact condition were in phase with body sway; subjects were essentially leaning on the contact surface through their finger for support. With light touch contact, CO[P.sub.x]-[F.sub.L] correlations were lower ([rho][approximately equal to].8), with time lags of approximately 300 milliseconds. Figure 3b shows that the increased time lag corresponds to the [F.sub.L] signal leading CO[P.sub.x] displacement. As subjects swayed toward the touch bar with only very light touch, contact forces initially increased, but as sway continued, [F.sub.L] decreased so as not to trigger the alarm threshold. This 300-millisecond lead of the force signal was maintained throughout the trial. The key point is that the additional stabilization provided by touch contact is due to a different sensorimotor relationship than with force contact. Forces generated by the musculature remote from the fingertip (eg, lower extremities, trunk) are guided by sensory information provided by cutaneous receptors in the fingertip(18,19) and proprioceptive information about arm position.(20,21)

The same infuence of light touch contact on postural control is observed in patients with bilateral loss of vestibular function. Patients with vestibular loss are thought to rely heavily on somatosensory information to indirectly derive missing information about head movement through the head-trunk linkage.(22) Consequently, an additional somatosensory reference from the fingertip may provide information about trunk movement that may enhance head-trunk coordination. In an unpublished study by Jeka and colleagues, five patients with complete bilateral vestibular loss and five age-matched subjects with normal vestibular function participated in the same paradigm as the above experiment. The subjects with vestibular loss were generally not able to maintain the tandem stance in the eyes closed-no contact condition for more than 5 seconds before falling. With light touch or force contact, however, postural sway was reduced to equivalent levels in all subjects. The same timing relationships between body sway and fingertip forces were observed in both the subjects with vestibular loss and the control subjects. Fingertip contact clearly substitutes for the sensory information that patients with vestibular loss lack to maintain upright tandem stance with eyes closed.

Leg Muscle Electromyographic Activity

In another experiment, Jeka and colleagues measured electromyographic (EMG) activity in the peroneal muscles, which are particularly important in stabilizing lateral body sway, to determine whether leg muscle activity changed with different contact cues. Two pieces of evidence suggest that light touch forces at the fingertip trigger EMG activity of the postural musculature. First, the EMG amplitude was lowest with force contact, higher with touch contact, and highest with no contact, indicating that leg muscles played a much larger role in maintaining balance with touch contact than with force contact. The increase in muscle activity with touch contact indicates that postural sway is reduced by additional muscular forces in the legs, whereas with force contact, the leg muscles play a lesser role. Forces generated by the arm musculature are a likely candidate to reduce body sway with force contact at the fingertip.

Second, shifts in the relative timing among CO[P.sub.x] displacement, contact force amplitude, and EMG activity were evident in the light touch versus force touch conditions. How the peroneal muscle's EMG activity behaved in each leg as the body moved in the mediolateral direction is illustrated in Figure 4. Electromyographic activity of the left peroneal muscle increased with CO[P.sub.x] displacement to the right (positive slope) and decreased with leftward CO[P.sub.x] displacement (negative slope). Right peroneal muscle EMG activity was approximately 180 degrees out of phase with left peroneal muscle EMG activity. In each condition, EMG activity led CO[P.sub.x] displacement by about 150 milliseconds. Together with the timing relationships between CO[P.sub.x] displacement and [F.sub.L] (Fig. 3), this finding means that with touch contact, changes in [F.sub.L] began about 150 milliseconds ahead of correlated changes in EMG activity. The observed pattern of directional changes in [F.sub.L] followed by leg muscle activity followed by body sway suggests that [F.sub.L] provided a feed-forward signal of body sway. In contrast, contact force changes were in phase with CO[P.sub.x] displacement in the force contact conditions, indicating that changes in leg muscle EMG activity were well ahead of changes in both CO[P.sub.x] displacement and fingertip contact force. This finding suggests that fingertip contact forces in the force contact conditions were not precuing a particular muscle activity pattern but were attenuating body sway primarily with physically supportive forces.

Touch Contact With a Cane

A recent study(15) compared the use of light touch forces between individuals with congenital blindness and sighted (eyes closed) individuals. A previous study(23) had shown higher levels of sway in blind versus blindfolded sighted individuals. A possible explanation may be that visual experience is a prerequisite for establishing a precise frame of reference for spatial tasks based on nonvisual information.(24-26) Moreover, the long cane is a commonly used mobility aid for blind individuals, and its primary function is thought to be obstacle avoidance.(27,28) Somatosensory cues from a cane also may provide a spatial referent that blind individuals use to stabilize upright stance.

Figure 5 shows the experimental setup. The touch bar was mounted to the wooden platform that rests on the force platform. The tip of a lightweight adjustable metal cane rested in a tiny well mounted to the touch bar so that the cane tip could not slide horizontally. Subjects were tested with the same three contact conditions as with the fingertip (ie, no contact, touch contact, and force contact). In addition, the touch bar was anchored in two positions: 60 cm lateral to the subject so that the cane was held perpendicular to the ground or 120 cm lateral to the subject so that the cane was slanted at a 30-degree angle relative to the ground. In each condition, cane length was adjusted so that the subject's elbow was flexed at 15 degrees. The five conditions were no contact (N), touch contact-perpendicular cane (TP), touch contact-slanted cane (TS), force contact-perpendicular cane (FP), and force contact-slanted cane (FS). Figure 6 shows the mean CO[P.sub.x] displacement results collapsed across five subjects with congenital blindness and five control subjects (sighted with eyes closed). Mean CO[P.sub.x] displacement was highest with no contact. Similar to the fingertip experiment, touch contact was as effective as force contact in reducing postural sway (compare conditions TP and FP or conditions TS and Fs). Moreover, postural sway was reduced most effectively with a slanted cane (conditions TS and FS), indicating that cane angle is an important consideration in its stabilizing influence (see "Light Touch With a Cane" section).

Timing Relationships With a Cane

Timing relationships between CO[P.sub.x] displacement and lateral cane force (C.sub.L) changed not only with the level of applied force, as in the fingertip studies, but also with cane angle. Mean CO[P.sub.x]-[C.sub.L] correlations and time lags in Figure 7 show negative correlations with time lags close to zero with a perpendicular cane (conditions TP and FP). With a slanted cane (conditions TS and FS), however, CO[P.sub.x]-[C.sub.L] correlations were positive, demonstrating that [C.sub.L] was now in phase with CO[P.sub.x] displacement. Using force contact with a slanted cane resulted in time delays close to zero, whereas touch contact with a slanted cane led to longer time delays (approximately 900 milliseconds), reflecting timing relationships very similar to those of the fingertip contact experiments (Fig. 3). Correlations between CO[P.sub.x]-displacement and vertical cane forces were positive and showed no differences due to cane angle. These results suggest that both vertical and lateral applied cane forces are involved in stabilizing postural sway and that stabilization is most effective when both directions of force are positively correlated to body sway.

How Does Light Touch Contact Reduce Body Sway?

These experiments have demonstrated that somatosensory contact cues at the fingertip and hand reduce postural sway in individuals without balance impairments, in persons with bilateral vestibular loss, and in individuals with congenital blindness. How do these "touch cues" serve as a source of sensory information about body orientation? Because muscular activity is greater in the peroneal muscles with light touch contact than with force contact suggests that light touch is triggering postural muscles to correct sway. The same peroneal muscles, however, are even more active without any contact. Why do we observe less sway with touch contact if muscle activity is higher with no contact? There are two possible explanations. One explanation is that additional postural musculature may be triggered with touch contact that is not active with no contact. The reference information provided by light touch contact may allow for a completely different set of muscles (eg, of the trunk) to counteract sway than with no contact, where somatosensory cues are derived primarily from around the feet and ankles. For instance, Winter et al(29) have shown that the hip abductors and adductors are prime candidates for the control of mediolateral body sway with feet side by side. Unfortunately, important hip abductor and adductor muscles, such as the gluteus medius, are deep muscles, making their EMG activity extremely difficult to isolate.

A second possibility for the reduction of body sway with light touch contact is the additional precision provided by somatosensory cues from the fingertip. Although cutaneous receptors are distributed across the entire body surface, they are particularly dense in the fingertips and hand. Analogous to the fovea of the retina, the fingertips are referred to as the "somesthetic macula."(30) Two-point discrimination studies have shown that the fingertip can resolve differences as small as 2 mm,(31) which is approximately the mean level of sway that we observe with light touch contact (Fig. 2). Interestingly, two-point discrimination at the bottom of the foot is approximately 8 to 10 mm, which is approximately the mean level of sway observed when subjects stand without fingertip contact and eyes closed.

Rapidly adapting (RA) cutaneous fibers, which have high spatial acuity and sensitivity to local vibration, are thought to be responsible for the detection of localized movement between the skin and a surface.(19) Slowly adapting (SA) cutaneous receptors, which are primarily responsible for tactual form and roughness perception through the distribution of forces across the skin surface.(19) may provide information about body sway through skin surface deformation or through "skin stretch."(32) In order to provide information about body orientation, however, cutaneous stimulation must be combined with knowledge of ongoing arm configuration that is dependent on interrelating muscle afferent and other proprioceptive activity to motor commands.(20,21,33)

There is much evidence to suggest that muscle spindle signals interpreted in relation to motor commands are the primary source of information for the position sense representation of the body and about body orientation relative to the support surface.(21,34) The fine acuity of cutaneous stimulation at the fingertip in combination with muscle spindle stimulation may allow for more precise detection of body sway with fingertip contact than somatosensory cues derived from the feet and ankles alone. Consequently, body sway is reduced because contact cues from the fingertip can detect trunk movements far earlier than those from the feet or ankles.

Light Touch With a Cane

Light touch through a cane had many of the same features of light touch with the fingertip. Jeka and colleagues' results suggest that the pattern of somatosensory stimulation with a hand-held cane could take two forms. All of the subjects they tested held the cane in the same manner, with palms resting on top of the cane and fingers wrapped around the handle in a thumb-opposing grip. The simplest pattern of stimulation could arise from body sway with the cane held still relative to the body. A traveling wave of stimulation across the palm may be interpreted as body sway, and appropriate muscular responses could inhibit further sway. Such somatosensory stimulation would be more difficult to interpret if the cane moved relative to the body. The pattern of stimulation could then be due to movement of the cane, movement of the body, or a combination of both.

As shown in Figure 8a, the negative correlation that Jeka and colleagues observed between body sway and lateral cane force with a perpendicular cane (Fig. 7) suggests that the cane and the body moved together around their respective pivot points (ie, cane tip and feet). In contrast, Figure 8b shows how positive correlations between body sway and lateral cane force suggest that the slanted cane was held still as the body swayed in the mediolateral direction. The strict threshold force of 1 N required in the light touch condition led subjects to adopt a sensorimotor relationship with a cane similar to that observed previously with the fingertip, that is, a similar lead of applied cane forces relative to body sway of 200 to 300 milliseconds. The reduction of sway that this relationship affords may be possible only with the high somatosensory precision of which the fingertips and hands are capable.

Clinical Implications

Clinicians often observe patients with balance disorders using light touch of surrounding objects and surfaces to stabilize themselves while standing and walking, but the actual use of touch contact or canes in balance control has not been studied systematically or rigorously. The findings of postural control with light touch contact may have potential applications to a large population of patients with balance and gait disorders due to neurological injury, including patients with hemiparesis, patients with Parkinson's disease, and elderly individuals. Such individuals are often capable of generating the appropriate muscular forces to maintain stable ambulation, but only if provided with sensory information that they are lacking due to neurological trauma. The additional sense of comfort that cane use provides may have little to do with physical support of the body. Two potential implications are addressed below: assistive device design and therapeutic mechanisms.

Assistive Device Design

The data on cane use with light touch contact have implications for assistive devices, where the primary design consideration is physical support for balance control.(28) Although rigidity and strength remain crucial features of any cane for instances when physical support is required, how often a cane is used for balance recovery (eg, after a "stumble") may be overestimated. Much of the time, I believe, a cane may serve the orientation function highlighted in the studies described, essentially fostering small corrections of upright stance that keep the center of mass well within stability limits and diminish the probability of a complete loss of balance. From this perspective, design features that enhance somatosensory feedback at the cane handle warrant further attention. These features may include 1) a texture of the surface at the handle that maximizes tactile resolution, 2) a shape of the handle that maximizes surface contact with the hand, and 3) indentations in the handle for the fingertips to maximize use of their fine spatial resolution.

Data indicate that the angle at which the cane contacts the ground relative to the body is a crucial influence on balance stabilization.(15) Although it is tempting to suggest that canes should be designed in a slanted manner to maximize their potential as a balance aid, a slanted shaft must be considered cautiously. Interviews of elderly cane users revealed that 30% expressed apprehension that their canes would cause them to fall as a result of tripping or a slippery tip.(2) Designers of mobility aids must recognize that the user is constantly maneuvering around obstacles that are not present in a laboratory or clinical setting. A cane with a slanted shaft increases the surrounding area that the user needs to navigate around obstacles. Thus, even if a slanted cane leads to better balance control than a vertical cane, the amount of angle may be limited by practical concerns in the real world. After extensive training, blind individuals manage to locomote freely with a cane that is extended further outward (ie, a long cane) than with a cane held vertically, but their primary goal is to perceive and avoid obstacles with the cane, even though some additional balance control may be derived. A possible compromise is to design a cane with a hinge joint near the end of the shaft that can be locked at a specific angle. This feature may afford the additional support provided by a small slanted section of the shaft while retaining the primarily vertical orientation of the shaft. Without testing these design ideas in the laboratory as well as the real world, it is possible only to speculate on their potential.

Therapeutic Mechanisms: Central and Peripheral

The inherent limitation to obtaining additional somatosensory information through the fingers or hand is that surfaces must be within reach. Obviously, this is not always practical. Patients with balance disorders often rely on light touch of a spouse for orienting information while walking through an environment without surfaces amenable to contact. These patients frequently dislike using a cane due to the image that it projects (eg, frailty). The main message from these studies of light touch is that somatosensory cues are useful for balance control up to the limits of resolving force changes across the skin surface. The fingertips have the most precise spatial resolution of any part of the body (along with the tongue) and, as a result, enhance balance control. The key for patients with balance disorders may be to artificially enhance the somatosensory information at the feet so that balance control is improved without the constraint of additional contact of surfaces with the fingertips or hand.

The potential for long-term restitution of function balance control through the enhancement of somatosensory cues at the feet may take advantage of central reorganization of cortical maps. Merzenich and Kaas(35) have shown that the body map in the somatosensory cortex of primates is capable of extensive changes following peripheral injury. Neurons in these central maps serving the damaged finger or area of the hand previous to injury reorganize to respond to adjacent areas of the skin surface. One explanation of such reorganization is that the inhibitory inputs from injured areas are eliminated, allowing previously suppressed inputs in adjacent intact regions to emerge as new receptive fields.(36,37) The time scale of reorganization is far more rapid than previously thought, on the order of hours after lesioning.(38)

Even without injury, however, a simple change of functional use can also result in rapid central reorganization. For example, primates trained on a retrieval task requiring skilled use of individual digits showed expansion of representation in the primary motor cortex for the fingers, whereas wrist and forearm zones contracted. In a second task involving forearm movements, forearm representation expanded, whereas digit representation contracted.(39) These results argue that differences in the cortical map structure across individuals are the consequence of differences in the functional use of peripheral limbs and that these changes are reversible. Such findings of neural plasticity based on functional use are being applied to rehabilitate motor and sensory function in patients with cerebrovascular accidents of the somatosensory cortex.(40) The therapy emphasizes tools that enhance sensory appreciation during retraining of tasks involving tactile form recognition and exploration. A similar approach may be feasible to enhance somatosensory input from the feet and ankles in elderly individuals and in persons with poor balance control.

The flip side of the message from these cortical reorganization studies emphasizes that central factors can potentially worsen a sensory deficit that begins as a peripheral problem. A person with early signs of peripheral neuropathy, for example, may decrease the reliance on a peripheral limb for everyday function. Such disuse could potentially trigger cortical reorganization that may hasten further deterioration of somatosensory function in the limb. Even though the time scale of cortical reorganization with a disease such as diabetic peripheral neuropathy may be slower than that of an acute injury such as a stroke, therapies designed to intervene at the first signs of peripheral deterioration may be most prudent.

Directions for Future Research

Studies of upright stance control in humans and animals have revealed that the postural control system reconfigures with changing sensory input. Nashner and Berthoz(41) demonstrated that initial muscle activation responses to a stabilized (ie, sway-referenced) visual surround were smaller in magnitude than with normal visual feedback. With successive trials, however, muscle activity regained levels equivalent to those with normal visual feedback, indicating that somatosensory and vestibular information were reweighted once subjects deemed visual information to be unreliable. Such reweighting is often characterized in terms of system gain.

Maioli and Poppele(42) studied the changes in functional limb length and limb orientation of standing cats to varying frequencies of tilted support-surface translations. They found that the ratio of the percentage of change in limb length to table tilt position remained constant (gain of about 1, in phase) to varying frequencies of support-surface translations. The authors suggested that input from an internal model of body orientation and dynamics adapts to the system's functional goals by selectively increasing gain.

Recent evidence has also shown adaptive increases in gain to a moving visual stimulus. Dijkstra et al(43) demonstrated that postural sway closely matched the amplitude of the visual motion even as distance to the visual display was varied. Quantitative modeling revealed that not only coupling strength to visual input but also the autonomous nonvisual component of the postural control system changed.(44) Using similar techniques, potential adaptive effects were observed when subjects used light touch contact of a moving surface,(45,46) illustrating that somatosensory information can drive postural sway similarly to full-field visual stimulation.

These studies suggest that the organizational scheme for upright posture is clearly not a fixed control system on which sensory information is imposed. Instead, the change in the underlying control system is not only based on a change in sensory information but also includes a change in the control system properties that are independent of the sensory stimuli.(47) Put most simply, these studies indicate that the sensitivity to a sensory stimulus can be selectively increased. The underlying mechanisms of adaptation in postural control are not well understood, particularly in patient populations and elderly individuals. For example, studies have shown that elderly individuals use inflexible postural control "strategies" that suggest a relatively fixed (nonadaptive) control system.(48) Can the inherently adaptive capability of the postural control system be used to develop better therapies? Can we rehabilitate individuals by artificially enhancing, for example, the sensitivity to somatosensation when vestibular information has been lost? Future studies should address this question if new rehabilitative techniques are to be developed.

Studies of light touch contact with a cane beg the question of whether these findings are applicable to dynamic balance activities such as cane use during ambulation. Some studies(5-8) have implemented force-sensitive canes to assess their aid to locomotion, but these studies recorded only the vertical direction of cane force and focused primarily on mean absolute levels of cane force. The results of the study by Murray et al(8) described earlier suggest that there is much more to using a cane than simple mechanical support, particularly across different patient populations. Their results, however, are primarily descriptive and provide no insight into the mechanisms or properties of cane use that lead to more stable locomotion (or more comfort as reported by the patients in their study). Therein lies a huge problem that remains to be solved. With so many interacting components, how does one characterize overall locomotor stability? The present results(13-15) (as well as those of Murray et al(8)) suggest that temporal relationships between limb/body movements and sensory input provide information that is crucial for investigations of human upright stance control. Temporal measures have been used to characterize stability of gait, such as cycle-to-cycle relative phase.(10) These measures require numerous cycles to estimate variability adequately and are often studied on a treadmill, where the relationship to overground locomotion remains speculative. Despite their difficulties, investigations of the coupling between sensory information and locomotory patterns have begun.(49)

In summary, a series of studies on postural control with light touch contact of the fingertip have demonstrated that somatosensory cues are a powerful orientation reference for improved control of upright stance. The movement of contact forces across the skin surface of remote extremities provides orientation cues about movement of the body and signals muscular activation for corrections of body sway. Small applied forces are not capable of physically moving the body, but they still provide information about body orientation relative to the surfaces on which we stand, lean, and touch. The improvement in balance control observed with a mobility aid such as a cane is often attributed to the cane acting as "third leg," with the concomitant widening of the base of support. Data suggest that in cases of a sensory deficit, improved balance control arises from the precise cues about body sway provided by somatosensory information from the fingertips and hand. The "third leg" is uniquely different from the real legs. It has the high resolution of the fingertip to detect force changes related to body sway, resulting in postural corrections well before the boundaries of upright stability are reached.

Whether individuals with balance problems actually use light touch with any regularity remains an open issue. Unfortunately, there is no evidence of patients spontaneously adopting a light touch strategy in the clinical environment. Personal communications with numerous physicians over the last few years, however, have led me to believe that light touch contact is often used for balance control. Reports of patients lightly touching their spouse's shoulder or arm during ambulation are common. The meaning of such an observation is difficult to assess because the actual force applied by the patient is never measured. One can, however, imagine situations in which it would be advantageous to use light touch contact rather than physical support of the body frame for balance control. For example, when using a railing for support while walking down a stairway, the more force that a person applies with the hand, the more frictional shear forces the person must overcome to move forward. A person may hold and release the railing with each step. With light touch contact, however, the frictional forces are so small that continuous contact is possible without inhibiting forward progress. The continuity of contact may result in a potentially safer strategy (ie, fewer falls). Even individuals without balance impairments experience advantages from light touch contact. When entering a dark room or corridor to search for a light switch, we often use light contact of furniture or objects to maintain balance when visual information is denied. Such observations remain speculative, however, until the actual use of light touch contact is studied in more detail and across a wider range of situations.

Acknowledgment

The valuable comments of Lisa DePasquale, PT, are gratefully acknowledged.

* In all of the studies reviewed here, center-of-pressure displacement was used to approximate overall body, sway. Center-of-pressure excursion is not equivalent to body, sway. Center-of-pressure excursions tend to be larger and of higher frequency than center-of-mass movements.(17) Pilot work, however, showed that correlations between center of pressure and a single light-emitting diode located at the navel and tracked with a video camera were found to average [rho]=0.9, with a 2.3-millisecond time lag. Such a high correlation is due primarily to the relatively small amplitude of overall body sway (< 1 cm) in the present paradigm. Thus, center-of-pressure displacement is assumed to be roughly equivalent to overall body sway.

References

1 Shoup TE, Fletcher LS, Merrill BR. Blomechanics of crutch locomotion. J Biomech. 1974;7:11-19.

2 Dean E, Ross J. Relationships among cane fitting, function, and falls. Phys Ther. 1993;73:494-500.

3 Joyce BM, Kirby Rib. Canes, crutches, and walkers. Am Fam Physician. 1991;43:535-542.

4 Milczarek JJ, Kirby RL, Harrison ER, MacLeod DA. Standard and four-footed canes: their effect on the standing balance of patients with hemiparesis. Arch Phys Med Rehabil. 1993;74:281-285.

5 Baxter ML, Allington RO, Koepke GH, Weight-distribution variables in the use of crutches and canes. Phys Ther. 1969;49:360-365.

6 Bennett L, Murray MP, Murphy EF, Sowell TT. Locomotion assistance through cane impulse. Bull Prosthet Res. 1979;10-31:37-47.

7 Robinson HS. Cane for measurement and recording of stress. Arch Phys Med Rehabil. 1969;50:457-459.

8 Murray MP, Seireg AH, Scholz RC. A survey of the time, magnitude, and orientation of forces applied to walking sticks by disabled men. Am J Phys Med. 1969;18:1-13.

9 Engel J, Amir A, Messer E, Caspi I. Walking cane designed to assist partial weight bearing. Arch Phys Med Rehabil. 1983;64:386-388.

10 Holt KG, Jeng SF, Ratcliffe R, Jamill J. Stability and the metabolic cost of human walking. In: Woollacott MH, Horak FB, eds. Posture and Gait: Control Mechanisms. Eugene, Ore: University of Oregon Press; 1992:392-395.

11 Pozzo T, Berthoz A, Lefort L. Head stabilization during various locomotor tasks in humans, I: normal subjects. Exp Brain Res. 1990;82: 97-106.

12 Pozzo T, Berthoz A, Lefort L, Vitte E. Head stabilization during various locomotor tasks in humans, II: patients with bilateral peripheral vestibular deficits. Exp Brain Res. 1991;85:208-217.

13 Jeka JJ, Lackner JR. Fingertip contact influences human postural control. Exp Brain Res. 1994;100:495-502.

14 Jeka JJ, Lackner JR. The role of haptic cues from rough and slippery surfaces in human postural control. Exp Brain Res. 1995;103:267-276.

15 Jeka JJ, Easton RD, Bentzen BL, Lackner JR. Haptic cues for postural control in sighted and blind individuals. Perception and Psychophysics. 1996;58:409-423.

16 Holden M, Ventura J, Lackner JR. Stabilization of posture by precision contact of the index finger. J Vestib Res. 1994;4:285-301.

17 Winter DA, Patla AE, Frank JS. Assessment of balance control in humans. Med Prog Technol. 1990;16:31-51.

18 Johansson RS. How is grasping modified by somatosensory input? In: Humphrey DR, Freund H-J, eds. Motor Control: Concepts and Issues. New York, NY: John Wiley & Sons Inc; 1991:331-355.

19 Johnson KO, Hsiao SS. Neural mechanisms of tactual form and texture perception. Annu Rey Neurosci. 1992;15:227-250.

20 Burgess PR, Wei JY, Clark FJ, Simon J. Signaling of kinesthetic information by peripheral sensory receptors. Annu Rev Neurosci. 1982; 5:171-187.

21 Matthews PBC. Proprioceptors and their contribution to somatosensory mapping: complex messages require complex processing. Can J Physiol Pharmacol. 1988;66:430-438.

22 Horak FB. Role of the vestibular system in postural control. In: Herdman SJ, ed. Vestibular Rehabilitation. Philadelphia, Pa: FA Davis Co; 1994:22-46.

23 Easton RD. Inherent problems of attempts to apply sonar and vibrotactile sensory aid technology to the perceptual needs of the blind. Optom Vis Sci. 1992;69:3-14.

24 Pick HL. Visual coding of non-visual spatial information. In: MacLeod RB, Pick HL, eds. Perception: Essays in Honor of James J Gibson. Ithaca, NY: Cornell University Press; 1974:153-165.

25 Rieser J, Guth D, Hill E. Mental processes mediating independent travel: implications for orientation and mobility. Journal of Visual Impairment and Blindness. 1982;76:213-218.

26 Rieser J, Guth D, Hill E. Sensitivity to perspective structure while walking without vision. Perception. 1986;15:173-188.

27 Blasch BB, Del'Aune WR. A computer profile of mobility coverage and a safety index. Journal of Visual Impairment and Blindness. 1992;86:249-254.

28 Farmer LW. Mobility devices. In: Welsh RL, Blasch BB, eds. Foundations of Orientation and Mobility. New York, NY: American Foundation for the Blind; 1980:357-412.

29 Winter DA, Prince F, Stergiou P, Powell C. Medial-lateral and anterior-posterior motor responses associated with centre of pressure changes in quiet standing. Neuroscience Research Communications. 1993; 12:141-148.

30 Phillips CG. Movements of the Hand. Liverpool, United Kingdom: Liverpool University Press; 1986.

31 Sherrick CE, Cholewiak RW. Cutaneous. sensitivity. In: Boff KR, Kaufman L, Thomas JP, eds. Handbook of Perception and Human Performance. New York, NY: John Wiley & Sons Inc; 1986:12-24.

32 Srinivasan MA, Whitehouse JM, LaMotte RH. Tactile detection of slip: surface microgeometry and peripheral neural codes. J Neurophysiol. 1990;63:1323-1332.

35 Matthews PBC. Evolving views on the internal operation and functional role of the muscle spindle. J Physiol (Lond). 1981;320:1-30.

34 Lackner JR. Some proprioceptive influences on the perceptual representation of body shape and orientation. Brain. 1988;111:281-297.

35 Merzenich MM, Kaas HH. Reorganization of mammalian somatosensory cortex following peripheral nerve injury. Trends Neurosci. 1982;5:434-436.

36 Dykes RW. Central consequences of peripheral nerve injuries. Ann Plast Surg. 1984;13:412-422.

37 Jacobs KM, Donoghue JP. Reshaping the cortical motor map by unmasking latent intracortical connections. Science. 1991;251:944-947.

38 Donoghue JP, Suner S, Sanes JN. Dynamic organization of primary motor cortex output to target muscles in adult rats, II: rapid reorganization following motor nerve lesions. Exp Brain Res. 1990;79:492-503.

39 Nudo RJ, Milliken GW, Jenkins WM, Merzenich MM. Use-dependent alterations of movement representations in primary motor cortex of adult squirrel monkeys. J Neurosci. 1996;16:785-807.

40 Dannenbaum RM, Dykes RW. Sensory loss in the hand after sensory stroke: therapeutic rationale. Arch Phys Med Rehabil. 1988;69:833-839.

41 Nashner L, Berthoz A. Visual contribution to rapid motor responses during postural control. Brain Res. 1978;150:403-407.

42 Maioli C, Poppele RE. Parallel processing of multisensory information concerning self-motion. Exp Brain Res. 1991;87:119-125.

43 Dijkstra TMH, Schoner G, Gielen CCAM. Temporal stability of the action-perception cycle for postural control in a moving visual environment. Exp Brain Res. 1994;97:477-486.

44 Giese MA, Dijkstra TMH, Schoner G, Gielen CCAM. Identification of the nonlinear state space dynamics of the action-perception cycle for visually induced postural sway. Biol Cybern. 1996;74:427-437.

45 Jeka JJ, Schoner G, Lackner JR. Entrainment of postural sway to sinusoidal haptic cues. Society for Neuroscience Abstracts. 1994;20:336.

46 Jeka JJ, Schoner G, Dijkstra TMH, et al. Coupling of fingertip somatosensory information to head and body sway. Exp Brain Res. In press.

47 Schoner G. Dynamic theory of action-perception patterns: the time-before-contact paradigm. Human Movement Science. 1994;13:415-439.

48 Horak FB, Mirka A, Shupert CL. The role of peripheral vestibular disorders in postural dyscontrol in the elderly. In: Woollacott MH, Shumway-Cook A, eds. Development of Posture and Gait Across the Life Span. Columbia, SC: University of South Carolina Press; 1989:253-279.

49 Warren WH, Kay BA, Yilmaz EH. Visual control of posture during walking: functional specificity. J Exp Psychol Hum Percept Perform. 1996;22:818-838.

JJ Jeka, PhD, is Assistant Professor, Department of Kinesiology, College of Health and Human Performance, University of Maryland at College Park, Room 2359, HHP Bldg, College Park, MD 20742-2611 (USA) ([e-mail]).

Dr Jeka was supported by a National Institutes of Health postdoctoral fellowship (1 F32 NS09025-02), by National Aeronautics and Space Administration grant NAG9-515, and by a Graduate Research Board grant at the University of Maryland.